Identity Protection Identity Protection
Risikosignale, Risikopolicies und automatische Remediation fĂŒr kompromittierte oder verdĂ€chtige IdentitĂ€ten. Risk signals, risk policies, and automatic remediation for compromised or suspicious identities.
FĂŒr den vollen Funktionsumfang inklusive Risk Policies ist Microsoft Entra ID P2 erforderlich. Microsoft Entra ID P2 is required for the full feature set including risk policies.
Identity Protection bewertet sowohl einzelne Sign-ins als auch den aggregierten Benutzerzustand. Identity Protection evaluates both individual sign-ins and the aggregated user state.
VerdÀchtige AktivitÀten erkennen und Benutzer automatisch oder manuell remediieren. Detect suspicious activities and remediate users automatically or manually.
Ăberblick Overview
Identity Protection ist die risikobasierte Erkennungsschicht von Microsoft Entra. Sie sammelt Telemetrie aus Sign-ins, Threat Intelligence und weiteren Microsoft-Sicherheitsquellen, um kompromittierte Konten und verdÀchtige Anmeldungen sichtbar zu machen. Identity Protection is the risk-based detection layer of Microsoft Entra. It collects telemetry from sign-ins, threat intelligence, and other Microsoft security sources to expose compromised accounts and suspicious sign-ins.
Die wichtigsten Objekte sind Risky users, Risky sign-ins und Risk detections. Darauf bauen Richtlinien fĂŒr MFA, PasswortĂ€nderung, Blockierung und Untersuchungsprozesse auf. The key objects are risky users, risky sign-ins, and risk detections. Policies for MFA, password change, blocking, and investigations build on top of them.
Detailinformationen zu vielen Detection-Typen sowie automatische risikobasierte Richtlinien setzen Microsoft Entra ID P2 voraus. Detailed visibility into many detection types as well as automated risk-based policies require Microsoft Entra ID P2.
Risikostufen Risk levels
| StufeLevel | BedeutungMeaning | PraxisOperational meaning |
|---|---|---|
| None | Kein ermitteltes Risiko.No detected risk. | Normale CA-/MFA-Policies greifen, aber keine risikobasierte Aktion.Normal CA and MFA policies apply, but no risk-based action. |
| Low | Anomalie mit geringer Sicherheit.Low-confidence anomaly. | Geeignet fĂŒr Beobachtung, abgestufte MFA oder zusĂ€tzliche Analyse.Suitable for observation, step-up MFA, or additional analysis. |
| Medium | Mehrere moderate Indikatoren oder stĂ€rkere Heuristik.Multiple moderate indicators or stronger heuristics. | Oft guter Trigger fĂŒr MFA oder intensivere Untersuchung.Often a good trigger for MFA or deeper investigation. |
| High | Hohe Sicherheit, dass Konto oder Sitzung kompromittiert ist.High confidence that the account or session is compromised. | Typisch fĂŒr Block oder sichere PasswortĂ€nderung.Typical for block or secure password change. |
Niedrige Risiken altern nach lÀngerer Zeit aus; mittlere und hohe Risiken bleiben bestehen, bis sie remediated oder manuell bestÀtigt bzw. verworfen werden. Genau deshalb sollten High-Risk-Ergebnisse nicht liegen bleiben. Low risks age out after time; medium and high risks remain until they are remediated or manually confirmed or dismissed. That is exactly why high-risk results should never be left unattended.
Risikoerkennungen Risk detections
Die folgende Tabelle fĂŒhrt die gĂ€ngigen Microsoft-Entra-ID-Protection-Detections fĂŒr Benutzer- und Sign-in-Risiko vollstĂ€ndig zusammen. Einige Erkennungen liefern variable Risikostufen abhĂ€ngig von der SignalgĂŒte. The following table combines the full list of common Microsoft Entra ID Protection detections for user and sign-in risk. Some detections can surface at variable risk levels depending on signal quality.
| NameName | TypType | Risk levelRisk level | Real-time / OfflineReal-time / Offline | BeschreibungDescription |
|---|---|---|---|---|
| Activity from anonymous IP address | Sign-in | Medium-High | Offline | Defender for Cloud Apps erkennt AktivitÀt von anonymen Proxy-IP-Adressen.Defender for Cloud Apps identifies activity from anonymous proxy IP addresses. |
| Suspicious MFA authentication approval | Sign-in | High | Real-time | Ungewöhnliche Eigenschaften rund um eine MFA-Freigabe deuten auf MFA-Phishing hin.Unusual properties around an MFA approval indicate possible MFA phishing. |
| Additional risk detected (sign-in) | Sign-in | Variable | Real-time / Offline | Abstrakter Hinweis auf Premium-Detection ohne Detailanzeige in niedrigeren Lizenzen.Abstract pointer to a premium detection without details in lower licenses. |
| Admin confirmed user compromised | Sign-in | High | Offline | Ein Administrator hat den Benutzer explizit als kompromittiert markiert.An administrator explicitly marked the user as compromised. |
| Anomalous Token (sign-in) | Sign-in | Low-High | Real-time / Offline | Abnorme Token-Eigenschaften deuten auf Replay oder Missbrauch hin.Abnormal token properties suggest replay or misuse. |
| Anonymous IP address | Sign-in | Medium | Real-time | Anmeldung von Tor-/VPN-/Anonymisierungs-IP.Sign-in from a Tor, VPN, or anonymizing IP. |
| Atypical travel | Sign-in | Low-High | Offline | Zwei geografisch entfernte Anmeldungen in unrealistischem Zeitabstand.Two geographically distant sign-ins in an unrealistic time window. |
| Impossible travel | Sign-in | Medium-High | Offline | Defender for Cloud Apps erkennt unmögliche Reisebewegungen.Defender for Cloud Apps detects impossible travel patterns. |
| Malicious IP address | Sign-in | High | Offline | Anmeldung von IP mit bösartiger Reputation.Sign-in from an IP with malicious reputation. |
| Mass Access to Sensitive Files | Sign-in | Medium-High | Offline | Massenzugriff auf potenziell sensible Dateien in SharePoint/OneDrive.Mass access to potentially sensitive files in SharePoint or OneDrive. |
| Microsoft Entra threat intelligence (sign-in) | Sign-in | Medium-High | Real-time / Offline | MSTIC-/Threat-Intel-basierte Mustererkennung.MSTIC and threat-intelligence-based pattern detection. |
| New country | Sign-in | Medium | Offline | Anmeldung aus einem Land, das fĂŒr den Benutzer ungewöhnlich ist.Sign-in from a country that is unusual for the user. |
| Password spray | Sign-in | High | Real-time / Offline | Erfolgreicher Password-Spray-Angriff gegen den Benutzer.Successful password spray attack against the user. |
| Suspicious browser | Sign-in | Medium-High | Offline | VerdĂ€chtiges Browser-Verhalten ĂŒber mehrere Tenants und Regionen.Suspicious browser behavior across multiple tenants and regions. |
| Suspicious inbox forwarding | Sign-in | Medium-High | Offline | VerdÀchtige automatische Weiterleitung aus dem Postfach.Suspicious automatic forwarding in the mailbox. |
| Suspicious inbox manipulation rules | Sign-in | Medium-High | Offline | VerdÀchtige Regeln zum Verschieben/Löschen von Nachrichten.Suspicious rules that move or delete messages. |
| Token issuer anomaly | Sign-in | Medium-High | Offline | Ungewöhnlicher oder kompromittierter SAML-Token-Aussteller.Unusual or compromised SAML token issuer. |
| Unfamiliar sign-in properties | Sign-in | Low-High | Real-time | IP, ASN, GerÀt oder Browser weichen deutlich vom Normalbild ab.IP, ASN, device, or browser deviates significantly from the user's norm. |
| Verified threat actor IP | Sign-in | High | Real-time | Anmeldung von einer verifizierten Nation-State- oder Cybercrime-IP.Sign-in from a verified nation-state or cybercrime IP. |
| Additional risk detected (user) | User | Variable | Real-time / Offline | Abstrakter Premium-Hinweis ohne Detailanzeige.Abstract premium hint without detailed visibility. |
| Anomalous Token (user) | User | Low-High | Real-time / Offline | Token-Anomalien tragen zum Benutzer-Gesamtrisiko bei.Token anomalies contribute to overall user risk. |
| Anomalous user activity | User | Medium-High | Offline | Anormales Verwaltungs- oder Verzeichnisverhalten.Abnormal administrative or directory behavior. |
| Attacker in the Middle | User | High | Offline | Reverse-Proxy-Session mit kompromittierender Adversary-in-the-Middle-Indikation.Reverse proxy session indicating an adversary-in-the-middle attack. |
| Leaked credentials | User | High | Offline | Validierte Zugangsdaten tauchen in bekannten Datenlecks auf.Validated credentials appear in known leaks. |
| Microsoft Entra threat intelligence (user) | User | Medium-High | Real-time / Offline | Threat-Intel-Signale verdichten sich auf Benutzerkontext.Threat-intelligence signals converge at the user level. |
| Possible attempt to access Primary Refresh Token (PRT) | User | High | Offline | Hinweis auf PRT-Zugriffsversuch und mögliche laterale Bewegung.Indicates attempted PRT access and possible lateral movement. |
| Suspicious API Traffic | User | Medium-High | Offline | AuffÀllige Graph- oder Verzeichnisabfragen deuten auf Reconnaissance hin.Abnormal Graph or directory API usage suggests reconnaissance. |
| Suspicious sending patterns | User | Medium | Offline | VerdÀchtige E-Mail-Sendemuster.Suspicious email sending patterns. |
| User reported suspicious activity | User | High | Offline | Benutzer lehnt MFA ab und meldet die Anforderung als verdÀchtig.User denies MFA and reports the prompt as suspicious. |
Risikorichtlinien Risk policies
| RichtliniePolicy | Typischer TriggerTypical trigger | AktionAction | KommentarComment |
|---|---|---|---|
| Sign-in risk policySign-in risk policy | Medium oder High Sign-in Risk.Medium or high sign-in risk. | MFA oder Block.MFA or block. | Sichert verdÀchtige Anmeldungen direkt beim Zugriff.Protects suspicious sign-ins directly during access. |
| User risk policyUser risk policy | High User Risk.High user risk. | Secure password change oder Block.Secure password change or block. | Ideal zur Selbstremediation kompromittierter Benutzer.Ideal for self-remediation of compromised users. |
| MFA registration policy / campaignMFA registration policy / campaign | Benutzer noch nicht registriert.User not yet registered. | Registrierung von Methoden erzwingen oder bewerben.Require or campaign for method registration. | Nicht direkt Identity Protection, aber eng mit Remediation verbunden.Not directly Identity Protection, but closely tied to remediation. |
Konfiguration Configuration
Sign-in Risk Policy konfigurieren Configure sign-in risk policy
- Entra admin center â Protection â Identity Protection â Sign-in risk policy öffnen.Open Entra admin center â Protection â Identity Protection â Sign-in risk policy.
- Benutzergruppe definieren und Notfallkonten ausschlieĂen.Define the user group and exclude emergency accounts.
- Acceptable sign-in risk level festlegen, typischerweise Medium und High.Set the acceptable sign-in risk level, typically Medium and High.
- Control = Require MFA oder Block access festlegen.Set the control to Require MFA or Block access.
- Erst mit Pilotgruppe testen, dann tenantweit ausrollen.Test first with a pilot group, then roll out tenant-wide.
User Risk Policy konfigurieren Configure user risk policy
- Protection â Identity Protection â User risk policy öffnen.Open Protection â Identity Protection â User risk policy.
- Hochrisikobenutzer auswÀhlen oder alle Benutzer definieren.Select high-risk users or target all desired users.
- Aktion = Require password change bevorzugen, wenn SSPR und Writeback sauber funktionieren.Prefer the action Require password change when SSPR and writeback are working cleanly.
- FĂŒr privilegierte Konten kann Block access sinnvoller sein.For privileged accounts, Block access can be the better option.
Registrierungsreife sicherstellen Ensure registration readiness
- Combined Registration und geeignete Recovery-Methoden bereitstellen.Provide combined registration and suitable recovery methods.
- Bei Hybrid Password Writeback vor Aktivierung der User Risk Policy testen.In hybrid estates, test password writeback before enabling the user risk policy.
- Support-Prozess fĂŒr Benutzer mit blockiertem Zugriff definieren.Define a support process for users who get blocked.
Untersuchung Investigation
| BerichtReport | FrageQuestion | VorgehenApproach |
|---|---|---|
| Risky usersRisky users | Welche Konten sind aktuell gefĂ€hrdet?Which accounts are currently at risk? | Benutzerhistorie, letzte AktivitĂ€t, Remediation-Status prĂŒfen.Review user history, recent activity, and remediation state. |
| Risky sign-insRisky sign-ins | Welche konkrete Anmeldung war verdĂ€chtig?Which specific sign-in was suspicious? | IP, App, Client, GerĂ€t, Zeitpunkt und Resultat prĂŒfen.Inspect IP, app, client, device, time, and result. |
| Risk detectionsRisk detections | Welche Signale haben das Risiko erzeugt?Which signals produced the risk? | Detection-Typ, Risiko, Entstehungszeit und verwandte VorfÀlle analysieren.Analyze detection type, risk level, detection time, and related incidents. |
Gute Investigation korreliert Identity Protection stets mit Sign-in Logs, Defender-Signalen, Exchange-/SharePoint-AktivitÀt und gegebenenfalls Endpoint-Daten. Erst die Kombination ergibt ein belastbares Bild. Good investigation always correlates Identity Protection with sign-in logs, Defender signals, Exchange or SharePoint activity, and where available endpoint data. Only the combination creates a trustworthy picture.
Behebung Remediation
| AktionAction | Wann verwendenWhen to use | WirkungEffect |
|---|---|---|
| DismissDismiss | False Positive oder bereits extern behoben.False positive or already remediated elsewhere. | Entfernt den offenen Risikozustand.Removes the active risk state. |
| Confirm compromisedConfirm compromised | Klarer Incident bestĂ€tigt.A real incident is confirmed. | Erhöht das Vertrauen in die Kompromittierung und unterstĂŒtzt stĂ€rkere GegenmaĂnahmen.Increases confidence of compromise and supports stronger countermeasures. |
| Confirm safeConfirm safe | Sign-in oder Benutzer verifiziert legitim.The sign-in or user is verified as legitimate. | Markiert den Befund als bereinigt oder harmlos.Marks the finding as safe or cleared. |
| Reset passwordReset password | Verdacht auf gestohlene Credentials.Suspicion of stolen credentials. | Unterbricht gestohlene Passwörter und kann User Risk remediieren.Cuts off stolen passwords and can remediate user risk. |
Automatische Behebung Automatic remediation
- Sign-in Risk Policy + MFA ermöglicht unmittelbare Self-Remediation wÀhrend der Anmeldung.Sign-in risk policy plus MFA enables immediate self-remediation during sign-in.
- User Risk Policy + Secure Password Change behebt kompromittierte Benutzer selbstÀndig, sofern SSPR, PHS und Writeback passend konfiguriert sind.User risk policy plus secure password change remediates compromised users automatically when SSPR, PHS, and writeback are configured properly.
- Bei privilegierten Konten sollte Self-Remediation oft durch enges SOC-/IAM-Playbook ergÀnzt werden.For privileged accounts, self-remediation should often be supplemented by a stricter SOC or IAM playbook.
Automatische PasswortÀnderung ist nur dann sinnvoll, wenn Benutzerregistrierung, Support-Prozess und Hybrid-Writeback sauber funktionieren. Sonst entstehen Lockouts und Helpdesk-Spitzen. Automatic password change only makes sense when user registration, support processes, and hybrid writeback work cleanly. Otherwise you create lockouts and helpdesk spikes.
CA-Integration CA integration
Identity Protection und Conditional Access ergÀnzen sich: Risk Signale liefern den Kontext, CA setzt die technische Durchsetzung um. Besonders bewÀhrt sind risikobasierte MFA- und Block-Policies. Identity Protection and Conditional Access complement each other: risk signals supply the context, while CA enforces the control. Risk-based MFA and block policies work especially well.
- Sign-in Risk = Medium/High â MFA.Sign-in risk = medium/high â MFA.
- Sign-in Risk = High â Block access.Sign-in risk = high â block access.
- User Risk = High â Require password change.User risk = high â require password change.
- Admins separat modellieren und fĂŒr High-Risk hĂ€ufig blockieren statt nur MFA zu fordern.Model admins separately and often block high-risk access instead of only requiring MFA.
PowerShell & Graph PowerShell & Graph
Connect-MgGraph -Scopes "IdentityRiskyUser.ReadWrite.All","IdentityRiskEvent.Read.All"
# List risky users
Get-MgIdentityProtectionRiskyUser
# List risk detections
Get-MgIdentityProtectionRiskDetection
# Confirm a user as compromised
Confirm-MgIdentityProtectionRiskyUserCompromised -RiskyUserId "user@contoso.com"
# Dismiss risk for a user
Dismiss-MgIdentityProtectionRiskyUser -RiskyUserId "user@contoso.com"
GET https://graph.microsoft.com/v1.0/identityProtection/riskyUsers
GET https://graph.microsoft.com/v1.0/identityProtection/riskDetections
POST https://graph.microsoft.com/v1.0/identityProtection/riskyUsers/confirmCompromised
Content-Type: application/json
{
"userIds": [
"11111111-1111-1111-1111-111111111111"
]
}
POST https://graph.microsoft.com/v1.0/identityProtection/riskyUsers/dismiss
Content-Type: application/json
{
"userIds": [
"11111111-1111-1111-1111-111111111111"
]
}
Best Practices Best practices
- Mindestens eine Sign-in-Risk- und eine User-Risk-Policy definieren.Define at least one sign-in risk policy and one user risk policy.
- Trusted Named Locations sauber pflegen, damit Risikoentscheidungen prÀziser werden.Maintain trusted named locations carefully so risk decisions become more precise.
- Privilegierte Konten mit strengeren Schwellen und getrennten Playbooks behandeln.Handle privileged accounts with stricter thresholds and separate playbooks.
- Identity Protection-Daten tĂ€glich ĂŒberwachen und mit SOC-Workflows koppeln.Review Identity Protection data daily and connect it to SOC workflows.
- Risk detections nicht isoliert lesen, sondern immer mit Sign-in- und Endpoint-Daten korrelieren.Do not read risk detections in isolation; always correlate them with sign-in and endpoint data.
Komplette Risikoerkennungs-ReferenzComplete risk detection reference
Die folgende Referenz fĂŒhrt verbreitete Sign-in- und User-Risk-Detections zusammen. Microsoft erweitert Bezeichner und interne Heuristiken regelmĂ€Ăig; produktive Automatisierungen sollten deshalb unbekannte Werte tolerieren, aber High-Risk-Ereignisse immer in Incident-Prozesse heben.The following reference consolidates common sign-in and user-risk detections. Microsoft extends identifiers and internal heuristics regularly, so production automations should tolerate unknown valuesâbut always escalate high-risk events into incident processes.
| riskEventTyperiskEventType | TypType | TimingTiming | LevelLevel | StateState | BeschreibungDescription | MitigationMitigation |
|---|---|---|---|---|---|---|
| anonymizedIPAddress | signIn | realTime | medium | atRisk | Anmeldung ĂŒber bekannte Anonymisierungsdienste oder Proxy-Infrastruktur.Sign-in through known anonymization services or proxy infrastructure. | MFA erzwingen, IP-Reputation prĂŒfen, Benutzer validieren.Enforce MFA, validate IP reputation, and confirm the user. |
| adminConfirmedUserCompromised | signIn | offline | high | confirmedCompromised | Administrator hat die IdentitÀt explizit als kompromittiert markiert.An administrator explicitly marked the identity as compromised. | Sofort blockieren, Sessions widerrufen, Passwort/Methoden rotieren.Block immediately, revoke sessions, and rotate password and methods. |
| anomalousToken | signIn | realTime/offline | medium-high | atRisk | Token-Eigenschaften deuten auf Replay oder Missbrauch hin.Token characteristics indicate replay or misuse. | Sessions widerrufen, Token-Path untersuchen, AitM prĂŒfen.Revoke sessions, inspect the token path, and investigate AiTM indicators. |
| anomalousUserActivity | signIn | offline | medium | atRisk | Ungewohnte Verwaltungs- oder VerzeichnisaktivitÀt im Sign-in-Kontext.Unusual administrative or directory activity in the sign-in context. | Sign-in Logs und Directory Audits korrelieren.Correlate sign-in logs with directory audits. |
| investigationsThreatIntelligence | signIn | offline | medium-high | atRisk | Threat-Intel-gestĂŒtzte Untersuchung fand bösartige Muster.Threat-intelligence-backed investigation found malicious patterns. | IOC-Suche, Defender- und SOC-Playbooks auslösen.Run IOC hunting and trigger Defender and SOC playbooks. |
| impossibleTravel | signIn | offline | medium-high | atRisk | Geografisch unmöglicher Wechsel zwischen Anmeldungen.Geographically impossible change between sign-ins. | VPN, Split-Tunnel und Session-Overlap prĂŒfen.Check VPN, split tunnel, and session overlap. |
| maliciousIPAddress | signIn | offline | high | atRisk | Quell-IP hat bestÀtigte bösartige Reputation.Source IP has confirmed malicious reputation. | Sofort blockieren und betroffene Sitzung untersuchen.Block immediately and investigate the affected session. |
| malwareInfectedIPAddress | signIn | offline | high | atRisk | IP wurde mit Malware- oder Botnet-AktivitĂ€t korreliert.IP was correlated with malware or botnet activity. | Endpoint-Status prĂŒfen und Device isolieren.Check endpoint state and isolate the device. |
| mcasImpossibleTravel | signIn | offline | medium-high | atRisk | Defender for Cloud Apps meldet unmögliche Reisebewegungen.Defender for Cloud Apps reports impossible travel. | MCAS-Sitzungen, ISP-Wechsel und Token-Reuse analysieren.Analyze MCAS sessions, ISP changes, and token reuse. |
| mcasSuspiciousInboxManipulationRules | signIn | offline | medium-high | atRisk | VerdĂ€chtige Exchange-Regeln wurden erkannt.Suspicious Exchange rules were detected. | Mailbox-Regeln entfernen und Postfachrechte prĂŒfen.Remove mailbox rules and review mailbox permissions. |
| mcasSuspiciousInboxForwarding | signIn | offline | medium-high | atRisk | Unerwartete externe Weiterleitung im Postfach.Unexpected external forwarding in the mailbox. | Forwarding stoppen, OAuth-Apps und Delegationen prĂŒfen.Stop forwarding and inspect OAuth apps and delegations. |
| newCountry | signIn | offline | medium | atRisk | Anmeldung aus einem fĂŒr den Benutzer neuen Land.Sign-in from a country that is new for the user. | Reisevalidierung, Geo-Blocking und MFA anwenden.Validate travel, geo-blocking, and MFA. |
| passwordSpray | signIn | realTime/offline | high | atRisk | Erfolgreicher Password-Spray-Angriff gegen das Konto.Successful password spray attack against the account. | Passwort Ă€ndern, Smart Lockout prĂŒfen, IPs blockieren.Change the password, review smart lockout, and block IPs. |
| riskyIPAddress | signIn | realTime | medium | atRisk | IP ist verdÀchtig, aber nicht zwingend bestÀtigt bösartig.The IP is risky, but not necessarily confirmed malicious. | MFA, Benutzerkontakt und weitere Telemetrie heranziehen.Use MFA, user confirmation, and additional telemetry. |
| suspiciousBrowser | signIn | offline | medium-high | atRisk | Browser-Fingerprints oder Session-Muster wirken verdÀchtig.Browser fingerprints or session patterns look suspicious. | Token-Diebstahl und AitM-Proxy untersuchen.Investigate token theft and AiTM proxy behavior. |
| suspiciousAPITraffic | signIn | offline | medium-high | atRisk | AuffÀllige API-Zugriffe deuten auf Reconnaissance oder Abuse hin.Abnormal API access suggests reconnaissance or abuse. | Graph-Logs, App-Permissions und Rate-Muster analysieren.Analyze Graph logs, app permissions, and rate patterns. |
| suspiciousIPAddress | signIn | offline | medium | atRisk | Die IP weicht stark vom erwarteten Benutzerprofil ab.The IP deviates strongly from the expected user profile. | Mit Sign-in-Historie und GerÀtedaten abgleichen.Compare with sign-in history and device data. |
| tokenIssuerAnomaly | signIn | offline | medium-high | atRisk | Ungewöhnlicher oder kompromittierter Token-Aussteller.Unusual or compromised token issuer. | Federation, Signing-Zertifikate und Claims-Flow prĂŒfen.Inspect federation, signing certificates, and claims flow. |
| unfamiliarFeatures | signIn | realTime | low-medium | atRisk | GerÀt, Browser, ASN oder Region weichen vom Normalbild ab.Device, browser, ASN, or region deviates from the normal pattern. | Step-up MFA und BenutzerbestÀtigung auslösen.Trigger step-up MFA and user validation. |
| verifiedThreatActorIPAddress | signIn | realTime | high | atRisk | Sign-in von einer bestÀtigten Nation-State- oder Crime-IP.Sign-in from a confirmed nation-state or crime IP. | Blockieren, IOC-Hunt starten, Incident hochpriorisieren.Block, launch IOC hunting, and prioritize the incident. |
| massAccessToSensitiveFiles | signIn | offline | medium-high | atRisk | Massenzugriff auf sensible Dateien in M365.Mass access to sensitive files in M365. | DLP, MCAS, SharePoint-AktivitĂ€t und Downloads prĂŒfen.Review DLP, MCAS, SharePoint activity, and downloads. |
| userReportedSuspiciousActivity | signIn | offline | high | atRisk | Benutzer hat die MFA-Anforderung als verdÀchtig gemeldet.The user reported the MFA prompt as suspicious. | Konto isolieren, Sessions widerrufen, Passwort und Faktoren rotieren.Isolate the account, revoke sessions, and rotate password and factors. |
| attackerInTheMiddle | signIn | offline | high | confirmedCompromised | Hinweise auf Reverse-Proxy- oder Adversary-in-the-Middle-Angriff.Indicators of a reverse-proxy or adversary-in-the-middle attack. | Phishing-Kit untersuchen, Tokens widerrufen, FIDO2 einfĂŒhren.Investigate the phishing kit, revoke tokens, and introduce FIDO2. |
| suspiciousMfaApproval | signIn | realTime | high | atRisk | Freigabe passt zeitlich oder technisch nicht zum Anmeldeversuch.The approval does not fit the sign-in attempt in timing or technical details. | MFA-Fatigue-Playbook ausfĂŒhren.Run the MFA-fatigue playbook. |
| unknownFutureValue | signIn | varies | unknown | atRisk | Platzhalter fĂŒr neue Detection-Typen in der API.Placeholder for new detection types in the API. | Automatisierungen tolerant auf unbekannte Werte auslegen.Make automations tolerant of unknown values. |
| leakedCredentials | user | offline | high | atRisk | Benutzerkennwort wurde in einem Leak gefunden.The user password was found in a leak. | Secure Password Change oder Block und Passwortrotation.Use secure password change or block and rotate the password. |
| userRiskAlerted | user | offline | medium-high | atRisk | Benutzer wurde als riskant markiert und alarmiert.The user was flagged as risky and alerted. | Incident-Triage und Benutzerkontakt starten.Start incident triage and user outreach. |
| additionalRiskDetected | user | realTime/offline | variable | atRisk | Premium- oder interne Detection ohne Detailanzeige in allen Lizenzen.Premium or internal detection without full detail in all licenses. | Mit Sign-in, Defender und Auditdaten korrelieren.Correlate with sign-in, Defender, and audit data. |
| possibleAttemptToAccessPrimaryRefreshToken | user | offline | high | atRisk | Hinweis auf PRT-Abgriff oder Credential Theft auf dem GerĂ€t.Indicates PRT theft or credential theft on the device. | Endpoint forensisch prĂŒfen und GerĂ€t isolieren.Inspect the endpoint forensically and isolate the device. |
| suspiciousSendingPatterns | user | offline | medium | atRisk | AuffĂ€lliges Sendeverhalten deutet auf Mail-Kompromittierung.Abnormal sending behavior suggests mail compromise. | Postfach prĂŒfen, Weiterleitungen entfernen, Benutzer sperren.Inspect the mailbox, remove forwarding, and block the user. |
| anomalousUserActivity | user | offline | medium-high | atRisk | Ungewöhnliche Directory- oder AdminaktivitÀt auf Benutzerebene.Unusual directory or admin activity at user level. | Role Usage, PIM und Directory Audits korrelieren.Correlate role usage with PIM and directory audits. |
| investigationsThreatIntelligence | user | offline | medium-high | atRisk | Threat-Intel-Signale verdichten sich auf denselben Benutzer.Threat-intelligence signals converge on the same user. | SOC-Untersuchung mit umfassender Telemetrie starten.Start SOC investigation with broad telemetry. |
| tokenIssuerAnomaly | user | offline | medium-high | atRisk | Ungewöhnliche Federation- oder Token-Issuer-Muster beeinflussen den Benutzerstatus.Unusual federation or token issuer patterns affect the user state. | Federation Trust und App-SSO prĂŒfen.Inspect federation trust and app SSO. |
| unknownFutureValue | user | varies | unknown | atRisk | Reservierter Wert fĂŒr zukĂŒnftige User-Risk-Detections.Reserved value for future user-risk detections. | Parser und Workflows robust auf unbekannte Werte bauen.Build parsers and workflows to tolerate unknown values. |
Workload Identity ProtectionWorkload identity protection
Workload Identity Protection ĂŒbertrĂ€gt Identity-Protection-Prinzipien auf Service Principals und App-only-Workloads. Im Fokus stehen kompromittierte Secrets oder Zertifikate, anormale API-Nutzung, verdĂ€chtige Sign-ins und ĂŒberprivilegierte Anwendungen.Workload Identity Protection applies Identity Protection principles to service principals and app-only workloads. It focuses on compromised secrets or certificates, anomalous API use, suspicious sign-ins, and overprivileged applications.
| DetectionDetection | BeschreibungDescription | ResponseResponse |
|---|---|---|
| Anomalous service principal activity | Ungewöhnliche Token-Ausgaben, API-Aufrufe oder Permission-Nutzung eines Service Principals.Unusual token issuance, API calls, or permission use by a service principal. | Secret/Certificate rotieren, Consent prĂŒfen, App ggf. deaktivieren.Rotate the secret or certificate, inspect consent, and disable the app if required. |
| Suspicious sign-in from service principal | App-only-Anmeldung aus ungewohntem Netzwerk, Tenant oder Laufzeitkontext.App-only sign-in from an unusual network, tenant, or runtime context. | Hosting-Standort prĂŒfen, Managed Identity bevorzugen, CA fĂŒr Workload Identities nutzen.Validate the hosting location, prefer managed identities, and use CA for workload identities. |
| Leaked credentials for service principal | Geheime App-Credentials sind kompromittiert oder öffentlich geworden.App credentials were compromised or became public. | Credential sofort ersetzen und alte Version widerrufen.Replace the credential immediately and revoke the old one. |
| Excessive privilege usage | Die App nutzt weit mehr Berechtigungen als betrieblich erwartet.The app uses far more permissions than operationally expected. | Least Privilege herstellen, App Governance und Reviews einfĂŒhren.Re-establish least privilege and introduce app governance and reviews. |
Connect-MgGraph -Scopes "IdentityRiskyServicePrincipal.Read.All","Application.Read.All"
Invoke-MgGraphRequest -Method GET -Uri "https://graph.microsoft.com/beta/identityProtection/riskyServicePrincipals"
Invoke-MgGraphRequest -Method GET -Uri "https://graph.microsoft.com/beta/identityProtection/servicePrincipalRiskDetections"Risikobasierte Conditional-Access-RezepteRisk-based Conditional Access recipes
Die stĂ€rksten Designs kombinieren Identity Protection mit klar getrennten Benutzerpopulationen. Die folgenden zehn Rezepte sind als Betriebsbausteine gedacht: mit Report-only beginnen, Ausnahmen dokumentieren und fĂŒr Administratoren immer eine schĂ€rfere Variante vorsehen.The strongest designs combine Identity Protection with clearly separated user populations. The following ten recipes are intended as operational building blocks: start in report-only, document exceptions, and always maintain a stricter variant for administrators.
| ## | KonfigurationConfiguration | BedingungenConditions | Grant / AktionGrant / action |
|---|---|---|---|
| 1 | Alle Benutzer â alle Apps â Legacy auth blockieren.All users â all apps â block legacy auth. | Client apps = legacy authentication clientsClient apps = legacy authentication clients | Block accessBlock access |
| 2 | Admins â Azure Management + Admin Portals â Phishing-resistant MFA.Admins â Azure Management + admin portals â phishing-resistant MFA. | Directory roles + Any locationDirectory roles + any location | Authentication strength = Phishing-resistant MFAAuthentication strength = phishing-resistant MFA |
| 3 | Alle Benutzer â Office 365 â MFA.All users â Office 365 â MFA. | Exclude break-glass + service exceptionsExclude break-glass + service exceptions | Require MFARequire MFA |
| 4 | BYOD mobil â Office 365 â App Protection erzwingen.BYOD mobile â Office 365 â enforce app protection. | Platforms = iOS, AndroidPlatforms = iOS, Android | Require approved client app OR app protectionRequire approved client app OR app protection |
| 5 | Unmanaged Browser â SharePoint Online â Download einschrĂ€nken.Unmanaged browser â SharePoint Online â restrict downloads. | Browser + Device state unmanagedBrowser + unmanaged device state | Use app enforced restrictionsUse app enforced restrictions |
| 6 | Sign-in risk medium/high â alle Benutzer â MFA.Sign-in risk medium/high â all users â MFA. | Sign-in risk = medium, highSign-in risk = medium, high | Require MFARequire MFA |
| 7 | Sign-in risk high â privilegierte Konten â Block.Sign-in risk high â privileged accounts â block. | Separate admin groupSeparate admin group | Block accessBlock access |
| 8 | User risk high â Mitarbeiter â Secure password change.User risk high â employees â secure password change. | User risk = high + SSPR readyUser risk = high + SSPR ready | Require password changeRequire password change |
| 9 | Guests â Microsoft 365 â MFA + Terms of Use.Guests â Microsoft 365 â MFA + terms of use. | Guests/external usersGuests/external users | Require MFA + Terms of UseRequire MFA + terms of use |
| 10 | Workload identity risk high â kritische Service Principals â Block oder Credential-Rotation-Playbook.Workload identity risk high â critical service principals â block or credential-rotation playbook. | Service principal risk = highService principal risk = high | Conditional Access for workload identities + automationConditional Access for workload identities + automation |